Key points are not available for this paper at this time.
Integrating proteomic and metabolomic data remains challenging due to the many-to-many relationships between metabolites and proteins and the spatial constraints of cellular compartmentalization. To address this, we developed PMconv, a web-based application for bidirectional, knowledge-based mapping of proteomic and metabolomic datasets. Leveraging curated associations from the Human Metabolome Database (HMDB) and protein interaction data from STRING, PMconv infers potential biochemical connections between experimentally detected molecules and pathway-annotated partners. The tool supports interactive network visualization and exports compartment annotations from the Human Protein Atlas to facilitate spatial contextualization of inferred interactions. PMconv is designed as an exploratory resource for hypothesis generation and feature engineering in multi-omics research, with the explicit understanding that knowledge-derived associations require experimental validation for compartment-specific interpretation.
Kozlova et al. (Thu,) studied this question.